For those of my readers going to PAS in Washington DC, I have put together a panel to discuss painful procedures in neonatal research projects, which starts at 2pm on Saturday, April 29, 2023. The stellar cast of speakers include the moderator John Lantos, who will introduce the topic and talk about research ethics and how painful interventions in children and babies fit into the ethical framework. I will give some examples of painful procedures in neonatal research projects, how they were justified, if at all, and try to create some guidelines for what might be acceptable, Ruth Grunau will talk about how pain adversely affects the developing brain, and Sunny Juul will talk about designing research projects that minimize additional pain.
I think it will be an important and informative session, hopefully with enough chairs for everyone who wants to attend. It is in the convention center 145 AB.
The definition of the outcome was “one or more of: intraventricular hemorrhage of grade 3 or 4, cystic periventricular leukomalacia, post-hemorrhagic ventricular dilatation, cerebellar hemorrhage, and cerebral atrophy” which are all defined more precisely in the supplemental material. One quibble I have with the definitions is of grade 4 IVH, defined as “parenchymal haemorrhagic infarction visible in the periventricular white matter”, what we actually see is echodensity, and it is not possible to really ascribe a mechanism of injury to an ultrasound appearance. As an upcoming review article that I co-authored will point out, periventricular echodenisities are not all strongly associated with poor long term outcomes. The extent, location and whether uni- or bi-lateral are all important.
This study was large enough (n=1600) to show moderately large differences in the incidence of ultrasound abnormalities, 90% power to show a reduction from 34 to 26% among infants <28 weeks GA. Monitoring had to start within 6 hours of birth and continue for 72 hours, a treatment algorithm was followed to protocolise responses to low cerebral saturations.
The results in the 2 groups were, basically, identical.
My big question is whether this is a surprise, I think that low cerebral oxygen saturations are due to a large number of different processes, and that there is not a clear link between low cerebral oxygen saturation and severe IVH or PVL. Only 29% of the babies in the oximetry group had a change in management because of low saturations, so you would only expect the intervention to have any impact on that 1/3 of the babies. PVL may occur after an episode of shock, but the strongest correlation is with perinatal inflammatory disorders, like chorioamnionitis.
I also think it is a real possibility that less cerebral hypoxia could improve long term outcomes without necessarily changing brain ultrasound findings, and I really hope that after this great effort to perform a really important trial, there will be long term follow up of the survivors.
For now there is no clear evidence that routine cerebral NIRS, and responding to low cerebral oxygenation, improves short or long term clinical outcomes in the preterm.
Discussions with parents about the progress of their baby, when we have concerns about their survival, or poor prognosis, are not infrequent in the NICU. We may decide to continue as before, or to limit interventions, or to withdraw some, or all, life sustaining interventions. One of our fellows, Beatrice Boutillier had collated the results of a large number of such discussions in a French tertiary NICU, and investigated the outcomes, who died, who survived after treatment limitation or withdrawal decisions, and the long term outcomes of those who survived. Annie Janvier and I collaborated with Beatrice, and with Valérie Biran, the chief of the NICU at Robert Debré hospital, to publish the data from their NICU.
There is very little other such information in the literature. Although there are several publications about the proportions of deaths that follow end of life discussions, many of those have collected data only on babies who died, there is little about those who survive after a decision to limit or withdraw care.
Although each case is unique, we squeezed them into general categories, using a revision of the categories of Eduard Verhagen et al. In some, all ICU care was withdrawn, including assisted ventilation, this could either be because it was thought there was little chance of survival in an unstable infant (category C), or because the long term prognosis for an acceptable quality of life was thought unlikely (category D). There were some infants where there was a decision to not escalate care, such as deciding to not go to the OR, or to not start an inotrope, or sometimes just to withhold CPR, these were category B. Those with withholding some interventions or limiting care were the babies most likely to survive, 18/29. But even among those with withdrawal of life-sustaining interventions, 4/41 unstable, and 12/94 stable infants survived.
All of the babies who survived to discharge required some medical follow up, and for many there were multiple specialties involved. Of the 34 who survived to 2 years, we had functional outcome information on 32, 8 of whom had outcomes between functionally normal, up to functional limitations with likely supervised living in the future, the remaining 24 were predicted to need help with activities of daily living or to be partially or totally dependent. We used the Glasgow Outcomes Score-Extended, as adapted for children (which I think gives a much better picture of the functional abilities of the children than a label of “NDI”).
The results point out the uncertainties inherent in our practice; even unstable babies that you think will likely die, who are so sick that you decide to withdraw their life-sustaining interventions, may still survive. After doing neonatology for over 40 years now (I know, I don’t look that old) I find myself often humbled by seeing things I would have thought impossible. Recently I looked after a baby with a pH <7.00 for 12 hours, (the baby was on extreme support, the parents weren’t ready to withdraw LST, so we continued while assuring good analgesia), who survived and is doing OK after discharge.
I am sometimes tempted to give up trying to predict anything! But I think it is better just to be transparent and honest and be open about the uncertainties of life, and of neonatology. We must always recognize that we are lousy at predicting survival, and even worse at predicting long term functional disability. Even with all the tools at our disposal, it is rare that we can say anything definitive.
One thing that doesn’t seem to help parents is to try and develop predictive models which give a percentage likelihood of survival or of poor long term outcomes. Although it is a different time of life, a recent study is relevant McDonnell SM, Basir MA, Yan K, Liegl MN, Windschitl PD. Effect of Presenting Survival Information as Text or Pictograph During Periviable Birth Counseling: A Randomized, Controlled Trial. Journal of Pediatrics. 2023. This study, using a vignette of a threatened delivery at 22 weeks gestation, showed that it didn’t make any difference how outcome data were presented (text or pictogram) to treatment decisions, it also showed that it didn’t make any difference what data were presented! Whether a 30% or 60% survival probability were revealed to a randomly selected group of 1000 women of child-bearing age didn’t make any difference to decisions, it didn’t even make any difference to what the participants thought the chance of survival was; when told the baby had a 30% chance of survival the participants thought the baby had a 68% chance of survival. Participants, however, thought that even with palliative care the baby was likely to survive (median 58% chance of survival), which does show that the presentation of the information, no matter how it was done, did not lead to the respondents having accurate knowledge of the results of palliative care.
“These seemingly nihilistic study results are illuminating. Should we be spending our time trying to get better at sharing outcome probabilities with expectant parents if the probabilities aren’t what matter to them?”
I never give percentage survival figures, unless parents ask for them (which they almost never do), and, to be honest, I am not at all sure that the difference between 30% survival and 60% should make a difference in decision-making. You might as well say 50:50, and leave it at that for both cases. There is a reasonable chance of survival, and ICU care is probably worth a try, I would say, and most of the respondents in the article by McDonnell seem to say the same thing. It might be different if the chances were 1% vs 99%, but it is hard to think of a scenario which could be included in such a study where the chances would be 1% survival.
To return to the case series that we just published, some of the considerations in counselling in the NICU are similar. We can try and calculate survival rates, but each case is different, each individual baby can only be 100% survival or 0% survival, and parental (and physician) attitudes and values are probably more important to the outcomes of such decisions than percentages. We have to always remember that neonatologists are remarkably bad at predicting the future.
We should also remember that parents of surviving preterm infants almost never regret their life and death decisions. In another study that we published recently, Thivierge E, et al. Guilt and Regret Experienced by Parents of Children Born Extremely Preterm. Journal of Pediatrics. 2022, 113 of 248 parents of former preterm infants, seen in the follow up clinic, expressed some regrets about the NICU stay. None of them regretted decisions made to start or continue NICU care. We know from other data that decisional regret is more common after decisions to opt for palliative or comfort care alone. The regrets that parents did express were more associated with lack of self-care during the NICU stay, regrets about having a preterm baby (for which mothers often blamed themselves) and regrets related to their role as parents in the NICU. Addressing those issues could help to improve parental mental health after discharge.
I had thought it was well settled that imaging soon after rewarming was as good for predicting outcomes in HIE as later imaging, but it seems that the timing, and the methodology are still uncertain.
The diffusion weighted MR images were preprocessed using software from the FMRIB Software Library (FSL) and MRtrix3. Preprocessing began with extraction of the b0 images from the diffusion weighted images using MRtrix3, followed by N4 bias correction of the b0 images. Post-bias correction, b0 images then underwent brain extraction via FSL’s BET to remove non-brain tissue, and to create a binary mask of the brain. Eddy current distortion correction and volume-to-volume motion correction was performed via FSL’s eddy current and motion correction using five iterations with the following parameters: outlier replacement, spline interpolation, least-squares reconstruction resampling, with full-width half maximums = [10,5,0,0,0] for each iteration. Post-eddy current and motion correction, FA values were estimated using FSL’s DTI FIT. Voxelwise statistical analysis of the FA data was carried out using Tract Based Spatial Statistics, part of FSL.22-23 These analyses were corrected for multiple comparisons using Threshold-Free Cluster Enhancement (FSL v1.2).
You might as well write “we did some stuff and got some pretty images of white matter”, it would be just as understandable for me! (and I think the vast majority of readers). However, one clear limitation of this method is that they were only able to get the pretty images in 20 of the 41 babies. The babies also had Prechtl standardised General Movements Assessment at 3 to 4 months of age, and 3 of those 20 were lacking the normal “fidgety movements”, all 3 of whom developed Cerebral Palsy (GMFCS 1,2 and 4) and 2/17 with such movements also had CP (GMFCS 1). There seemed to be a correlation between the infants who had absent fidgety movements and DTI images with abnormal Fractional Anisotropy in the Corpus Callosum, Posterior Limb of the Internal Capsule, and thalamic and optic radiations. The same changes were also found in the babies with CP who did have fidgety movements. In the results it is not at all clear of there was an association between DTI findings and long term outcomes, there is a brief statement that the same regions were different between those with and without CP, but no other mention of the results of the Bayleys, that all the babies had. The study might give an indication of which tracts are involved in the abnormal motor development that leads to absence of fidgety movements, and has research relevance, but certainly not any reason to add to clinical imaging.
Garvey AA, et al. Differences between early and late MRI in infants with neonatal encephalopathy following therapeutic hypothermia. Pediatr Res. 2023. This is a review from Boston of the images and outcomes of 94 babies who had both early, <7 days, and late, >7 days, MRI. The early images were obtained at a median of 4 days (IQR 4-4, or in other words more than half of them on day 4) and the late ones at a median of 16 days (IQR 12- 25 days). The second scan was obtained in order to satisfy their protocol which was to have a 2nd scan at 10 to 21 days, and it seems most were obtained from babies who were still inpatients, there were a concurrent 156 babies who were cooled and only had one scan (2 died before, 148 were discharged before, and 4 transferred). Most of the normal scans (21/24) were normal at the later date, and most of the abnormal ones (59/70) were also abnormal later, with mostly very similar findings, although 4 were worse. Three of the normal scans were abnormal later, and 11 of the 70 abnormal scans were normal at the second scan. You can see a summary of the results in this table:
The authors of this study refer to the systematic review of Ouwehand S, et al. (Predictors of Outcomes in Hypoxic-Ischemic Encephalopathy following Hypothermia: A Meta-Analysis. Neonatology. 2020;117(4):411-27), which claimed that the early MRI was more predictive of outcomes (variable outcome depending on the study) than later images. But there are very few studies in that review that actually scanned the babies twice, most of the included studies used either early or late MRI, not both; and with the variability between cohorts, variability in outcomes analyzed, and variability in interpretation, I am not all sure that is a reliable conclusion. Standard MRI images may not show injury early, so some of that difference in predictive power may be because different sequences (diffusion weighted) will find abnormalities in early scans.
Parmentier CEJ, et al. Additional value of 3-month cranial MRI in infants with neonatal encephalopathy following perinatal asphyxia. J Pediatr. 2023:113402. Maybe we should just keep repeating the MRI? (facetious comment). This retrospective review of MRI repeated at 2 to 4 months, apparently they all had neonatal MRI abnormalities. The babies were also followed up to 18 to 24 months, with developmental testing and a neurological exam. (BSID III motor or cognitive score <85 considered abnormal). 41 of the infants were seen again at 5.5 years of age, and had further intellectual and neurological assessments. There was some evolution, as you might expect, between the neonatal and 3 months images, but as the authors themselves note, there seemed to be little additional prognostic value with the later scans. The scores derived from the 3 month MRI were not closely associated with 5.5 year outcomes, mostly not being significantly different between those with normal outcomes and those with abnormal outcomes, defined as a WPSSI IQ scores <85, and subscale scores of <85 or epilepsy, deafness, blindness, or CP.
Wu YW, et al. How well does neonatal neuroimaging correlate with neurodevelopmental outcomes in infants with hypoxic-ischemic encephalopathy? Pediatr Res. 2023. This is an analysis of data collected during the HEAL trial, of erythropoietin in addition to cooling in term HIE (compared to cooling alone). As you probably remember, the trial had a negative result, with no additional benefit of erythropoietin. As all the babies had MRI, and all had neurological and developmental follow up the trial gives a real world evaluation of the predictive capacity of MR imaging in a multicenter prospective trial. All the babies had T1 and T2 and Diffusion Weighted images, and MR spectroscopy of left thalamus. The main outcome of interest, I think, is what is referred to as “severe NDI”, which seems to have been BSID cognitive score <70 or a severe motor outcome, which is defined in a unique way, with a modified GMFCS, but using a matrix determined by the GMFCS and a diagnosis of CP, and its type. So an infant with a clinical diagnosis of spastic tetraparesis is considered a severe outcome even if the GMFCS is 1. An infant with hemi- or di-plegia is considered severe at a GMFCS of 3.
The scoring system for the MRI was one previously validated, which produced 3 grades of abnormality.
As this figure shows, there was no difference in the BSID scores between normal, mild or moderately abnormal 5 day MRI, only severely abnormal MRI was associated with much lower scores. The range of scores with normal or less severely abnormal MRI is worrying though, with some babies at each stage having very low scores, and some having normal or even high scores. Even the babies with severely abnormal MRI had a wide range, extending up to scores in the normal range. The MRS results show a similar pattern. The figure that I found most useful was the summary of the PPV, NPV, sensitivity and specificity of the various findings. Unfortunately they only produced this figure for death or “NDI”, which included BSID cognitive score <85, and mild motor dysfunction.
In this cohort who all had moderate or severe HIE, the PPV and NPV of almost all findings are poor. Clinical severity of the HIE is just as good as most findings on the MRI, apart from “Global Injury”. What I get from this is that the MRI score, and MRI severity of injury are quite poor at predicting medium term outcomes, but the pattern of injury, and specifically the presence of Global Injury (which is defined as damage affecting >75% of the brain, including Central Gray Matter, White Matter, and Cortex. Although not shown in this figure, in the text it states “with all 34 affected infants either dying or developing severe NDI”. The sensitivity for that finding is poor, as there were many infants with “NDI” who did not have global injury, and therefore the NPV is only around 50%, but the specificity and PPV are very high, and in this case, the outcomes are of major clinical importance.
I wonder about the clinical evolution in those babies; most of the babies I have seen who have the most extensive MRI injuries also have a very unfavourable clinical findings, with poor or absent feeding and seriously abnormal clinical exam.
I know I will come across as an “imaging nihilist”, but I really wonder about the added value of MRI imaging in these babies for our families. I know all the real experts will say that routine MRI, just after rewarming, with T1, T2, and Diffusion Weighted imaging is the “standard of care”.
The PPV and NPV of most findings, as shown in that last figure, are around 50%, so when you talk to parents after the MRI, all you can really say is with these findings, your baby has about a 50:50 chance of having some developmental delay, and some motor problems. Which is what you also have to say if the MRI is normal!
If you read the first vignette, then the clinicians predictions, you will see that almost all thought the child would have moderate or severe disability, and many were very or extremely confident in their predictions, although there was a wide spread in how much they thought the MRI helped them. In fact the child is doing OK, with an isolated language delay. In another vignette, the baby had a catastrophic start to life, and was extubated day 5 with ongoing severe encephalopathy and lack of a gag reflex, the MRI on day 10 showed relatively minor abnormalities, but the baby had a very poor long term outcome. Most clinicians felt the MRI didn’t help them very much (without knowing the long term result).
As for the parents responses, I would urge you to read the quotes, the only one that surprised me was one parent who found the doctors rude; knowing the gang at McGill fairly well, I can only assume that someone was having a bad day, and perhaps also that in this incredibly stressful environment, even a single misplaced word can have real impacts. The clinicians all felt they needed more training, both in how to interpret MRI scan reports and in how to communicate the results. As someone who trained before MRI was invented, (I remember the first magnetically isolated room being constructed in Edmonton) and who has gradually learned on the job without formal training, I can only echo those desires. The reports of the MRIs in the vignettes illustrate how they are somewhat opaque. I try to always discuss the MRI results alongside the neurologist, but I am much more uncertain about the predictive utility of the MRI than they tend to be.
Indeed, the results of the studies I have reviewed above confirm that we are right to be uncertain! I really question the added value of the MRI to families. The only finding which is reliable enough to make a definitive statement about prognosis is global injury, which mostly occurs in babies with severe clinical findings. All other grades, patterns, and extent of injury have very poor PPV and NPV. Before writing this I thought the NPV of a normal scan was good enough that we could be reassuring to parents, and that they could relax somewhat. But as the HEAL study shows, if you start off with moderate to severe HIE, even with a normal MRI about 20% of the babies will have moderate or severe problems with development or motor skills.
A long break since my last post. I have had a couple of trips, to Arizona and to Australia, and more recently to Washington DC. With each conference I now try, if I can, to take an extra few days to do some wildlife photography. You can see some of the results of the Arizona trip here, and the Australia trip here. Or head over to the home page of my personal blog and browse the many images that are now there.
The first of these recent publications that I will discuss here presents some of the data from the Parents’ Voices Project asking parents of extremely preterm infants in follow-up about their concerns. (Thivierge E, et al. Pulmonary important outcomes after extremely preterm birth: parental perspectives. Acta Paediatr. 2023) 44% of the nearly 300 respondents (all parents of extreme preterm infants) voiced concerns about pulmonary outcomes. Some of them related to the initial hospitalisation, and others about longer term respiratory difficulties. The concerns could be categorized as those related to how the babies were treated (such as, in the short term, duration of intubation and reintubations) and those reflecting the severity of the lung injury, such as needing oxygen at home, and the duration of oxygen treatment after 40 weeks PMA.
In the longer term there were also concerns relating to treatments received, such as the isolation of the family as they attempted to avoid infectious contacts, and indicators of lung injury, such as the impacts of increased respiratory distress (loss of sleep, need for respiratory medications). This table is a listing of the categories of concern that were developed from the answers to open-ended questions in a study conducted among parents of extremely preterm infants.
One thing that none of the parents mentioned was being in oxygen at 36 weeks, or a clinical diagnosis of bronchopulmonary dysplasia. Our focus on BPD, as diagnosed at 36 weeks, can therefore only be justified, in my view, if it is a good interim outcome, which strongly predicts long term respiratory pathology of interest to families. Unfortunately it does not do so.
With the simpler dichotomous definitions, whether or not you include the requirement for Chest X-ray abnormalities, the positive predictive value for these outcomes is poor, as you can see it is around 40% for medication use, and 36% for rehospitalisation. The negative predictive value is also very poor, 30% of babies without a diagnosis of BPD nevertheless were receiving diuretics or bronchodilators, and 25% were readmitted to hospital for respiratory problems. The consensus NICHD definition doesn’t fare much better, many babies without BPD have those outcomes, and more than 50% with severe BPD do not.
Another very recent publication compared different BPD definitions at 36 weeks with long term outcomes, among babies from a single NICU in Amsterdam (Katz TA, et al. Comparison of New Bronchopulmonary Dysplasia Definitions on Long-Term Outcomes in Preterm Infants. J Pediatr. 2023;253:86-93 e4).. In this study an adverse respiratory outcome was “defined as occurrence of at least one the following: (1) ≥3 hospitalizations for respiratory and nonrespiratory reasons,24 (2) a visit to a respiratory specialist, (3) oxygen use at home between discharge and follow-up visits, and (4) chronic use of bronchodilators, antibiotics, or diuretics at 2 consecutive follow-up visits.”. These morbidities occurred in 22% of the babies at 2 years, and 18% at 5 years.
For any of the definitions they evaluated, the Area Under the Curve of the ROCs was not much better than chance.
Graph “C” is the ROC for 2 year adverse respiratory outcome, and “D” for the 5 year outcome. The AUC (or C-statistic) were all between 0.55 and 0.58. Which is not much better than flipping a coin at 36 weeks! (AUC 0.5), and none of which were “statistically significant”. The ROC’s for Neurologic impairment or developmental delay were slightly better, but still mostly consistent with randomly predicting which baby would have difficulties.
I wonder what the best outcome would be for respiratory interventions in the preterm infant? In my mind the situation is similar to cerebral injury and neurodevelopmental outcomes, we don’t use head ultrasound findings as the primary outcome for brain protection studies, or for quality control. Although very imperfect, developmental evaluation at 2 years of age and neurological abnormalities at that age are what we use as the primary outcomes. Why not focus on respiratory outcomes during that same period as our primary measure of efficacy for respiratory interventions? These babies are all being followed anyway, so the additional cost would not be enormous, and most of the outcomes would be from parent questionnaire.
As an interim outcome, oxygen use at 40 weeks is a better predictor of serious long term morbidity than at 36 weeks, in the large CNN cohort study of Isayama T (Revisiting the Definition of Bronchopulmonary Dysplasia: Effect of Changing Panoply of Respiratory Support for Preterm Neonates. JAMA Pediatr. 2017;171(3):271-9), but in that study serious respiratory morbidity was quite serious! “(1) 3 or more rehospitalizations after NICU discharge owing to respiratory problems (infectious or noninfectious); (2) having a tracheostomy; (3) using respiratory monitoring or support devices at home such as an apnea monitor or pulse oximeter; and (4) being on home oxygen or continuous positive airway pressure at the time of assessment between 18 and 21 months corrected age”.
This outcome occurred in only 16% of infants on O2 or respiratory support at 40 weeks gestation, compared to 2% who were not on such support. In other words the PPV for that outcome was poor, but the NPV was very high. I would love to know what were the PPV and NPV for less extreme pulmonary morbidity. But at least oxygen at 40 weeks seems to be something that we could use as an interim outcome that parents actually care about: why do they care? I presume it is because it is a sign that the baby is staying in hospital because of their respiratory status, and their discharge is delayed as a result. It is quite uncommon in my NICU, and many others across Canada to send babies home on O2 before 40 weeks, they are usually at least at 42 weeks in my NICU before we organise home O2 if we discharge them with it.
Now that we have some sort of idea of the respiratory outcomes that matter to parents, we should try and find out how they prioritize them, and then try and construct an outcome variable that takes them into account, and that parents agree reflects the severity of their respiratory morbidity. Some recent papers have suggested that maybe we should do pulmonary function tests, which have the advantage of being relatively objective, but the major disadvantage of being somewhat invasive in young infants, relatively expensive, and not always correlating well with clinical impacts. I think a better idea of how respiratory morbidity affects the infant’s and family’s daily life would be preferable, and should be a priority for researchers in respiratory care of the preterm infant.
For another project I have been involved in, I reviewed a lot of studies addressing how and when preterm babies develop their feeding skills. Which I thought I would share.
Oral feeding is complicated! The first thing a baby has to do is to get milk from the breast into their mouth, this requires a lot more than suction. There needs to be a coordinated sequence of suction to keep the nipple in the mouth, and compression of the nipple by the tongue, coordinated movements of the palate and movement of the milk to the pharynx.
You can see the tongue as the vaguely tongue-coloured block in the middle compressing the nipple, helpfully enhanced with the white milk ducts running through it, and the light blue coloured milk being ejected, and then moved to the back of the mouth into the pharynx. The nasal airway (in darker blue) is protected by the tongue being in contact with the palate, which then closes off the nasal airway as the milk bolus passes. It is easy to imagine the developmental maturity required for all of that to pass off without a hitch, and that is just the beginning, of course.
The subsequent stages include the initiation of swallowing and its co-ordination with a respiratory pause, and adequate oesophageal function. These steps require pharyngeal contractions to propel the milk posteriorly, relaxation of the upper oesophageal sphincter to allow passage and then oesophageal peristalsis, all of which have to be coordinated with respiration. In this study from Jadcherla’s group in Columbus OH, the coordination of suction and pharyngeal contraction was investigated, after the administration of a small liquid bolus into the pharynx, (Hasenstab KA, et al. Maturation Modulates Pharyngeal-Stimulus Provoked Pharyngeal and Respiratory Rhythms in Human Infants. Dysphagia. 2018;33(1):63-75)
If you follow the alphabet, A the initial pharyngeal response was multiple pharyngeal contractions, during which time (C) respiration paused, and oesophageal activity ceased, then the baby took a breath, and more pharyngeal contractions (B) propelled the liquid to the oesphagus, where a peristaltic wave can be seen descending, as respiration has re-started.
It isn’t surprising that this entire sequence is not mature in the fetus, who doesn’t need to breathe, or in the very preterm infant, who shouldn’t need to be breathing. In the same publication Hasenstab and others recorded 18 preterm infants (median GA 27 weeks) when they reached term, and again 4 weeks later.
The following figure illustrates the changes between the 2 recordings: the initial apnea was substantially longer at term, than later, and there were more peaks in the pharyngeal pressure, which were more variable.
In another study of 48 preterm infants at term equivalent age (Hasenstab KA, et al. Pharyngoesophageal and cardiorespiratory interactions: potential implications for premature infants at risk of clinically significant cardiorespiratory events. Am J Physiol Gastrointest Liver Physiol. 2019;316(2):G304-G12), half of whom were extreme preterm, heart rate decelerations occurred more frequently after a pharyngeal stimulus (0.1 to 0.5 mL of saline infused through a pharyngeal catheter) in the more immature infants, but most had a heart rate acceleration. Most of these studies have been performed during a bottle feed, as measuring pressures and flows is much easier than during breastfeeding! We have to assume that much of the physiology of the swallow, respiration, oesophageal activation, phase is the same. There are a couple of studies that have collected physiologic data during breast feeding, from the Perth Australia group, who have measured intra-oral pressures and performed simultaneous ultrasounds. They are responsible, I think, for the ultrasound image at the top of this post, and have published a cros-sectional study of preterms
Geddes DT, et al. Characterisation of sucking dynamics of breastfeeding preterm infants: a cross sectional study. BMC Pregnancy Childbirth. 2017;17(1):386. Th3 38 infants studied were of various gestational ages, but had all progressed to be feeding at the breast and where around 35 weeks PMA. Overall they showed that babies at this postnatal stage were able to generate a vacuum and had tongue movements similar to term babies, but they generated much less negative pressure (-40mmHg compared to the -115 that term babies generate). The preterms also had very variable sucking patterns, and also tended to have a progressive decrease in sucking vacuum during a feed, whereas term babies usually have a progressive increase. Indeed, as the nurses are often telling us, preterms do get tired during feeds! Full term babies have multiple sucking bursts, of more than 10 sucks, and few isolated sucks, here again the preterm babies were different, with fewer good bursts of <10 sucks, and more isolated single sucks. Of note, almost all of these babies/mothers were using nipple shields, which are of uncertain value, they may perhaps improve the efficiency of breast feeding, but might have an adverse effect on breast feeding duration.
Kerstin Nyqvist in Uppsala has studied a group of 15 extremely preterm infants, born between 26 and 31 weeks gestation, who had a history of either assisted ventilation (median duration 1 day) or CPAP, who were either in room air or on low flow oxygen. (Nyqvist K. Early attainment of breastfeeding competence in very preterm infants. Acta Paediatr. 2008;97(6):776-81) Most of the observations published were recorded by the mothers, after being taught what to observe:
You can see that the first signs of rooting occurred as early as 29 and a half weeks in one baby, but sometimes as late as 34 weeks, and more prolonged fixation and longer sucking bursts occurred progressively later. Also there was marked variability in the PMA at which certain activities were documented.
Can we help babies to develop these skills earlier?
Another small (n=35) RCT from Marseille examined the use of a “uni-modal orofacial stimulation” technique, whereby the babies, 26 to 29 weeks GA, at <33 weeks PMA, had 12 minutes of stimulation of the cheeks, lips, gums and tongue by a gloved hand, followed by 3 minutes of non-nutritive sucking (presumably a soother), twice a day. Méziane S, et al. Cardio-Respiratory Events and Food Autonomy Responses to Early Uni-Modal Orofacial Stimulation in Very Premature Babies: A Randomized, Controlled Study. Children. 2021; 8(12). The primary outcome of the study was occurrence of apnoea and bradycardia episodes, but they also investigated the volume of milk taken during the first feed (presumably a bottle feed), which was 34mL compared to 29mL, and the subsequent feeds, which were not different between groups.
As far as I can tell, from this review, promoting successful breast feeding in extremely preterm infants should include introduction to the breast starting early during skin to skin care, that the infants will start to show an interest, and preliminary rooting at between 30 and 32 weeks, and they may initiate single sucks soon thereafter. Whether there is any intervention that can improve the success rate of breast feeding, and can lead to earlier fully established breastfeeding, with potentially earlier discharge and fewer long-term feeding problems, remains to be clearly shown.
The authors of the Beneductus trial commented on my post about the trial, but it has somehow disappeared from the comment section of the blog, as they raise very valid points, I thought I would copy what they sent here as a new post.
First, I would like to say that I often try to write my posts to be provocative, and hopefully to make us all consider how best to treat newborn infants. Although I might criticise how a trial was done, I have huge respect for everyone who tries to do prospective research. We will never advance if everyone sits about writing blogs about others research, rather than doing the very difficult work, which always involves multiple compromises, of performing new trials. That being said, here is the text of the comment from Willem de Boode and Tim Hundscheid about the trial.
Thank you for your interest in our recent publication entitled ‘Expectant Management or Early Ibuprofen for Patent Ductus Arteriosus’. (1)
We have read your blog, posted on December 21, 2022 with great interest and would like to comment on this.
The first question you raised was about the mortality rate in the two study groups, more specifically death before discharge. Although mentioned in Table S2 – Outcome parameters with definitions, the mortality prior to discharge is not reported in the paper. This is related to the fact that death before discharge was identical to the mortality at 36 weeks postmenstrual age. In the original Table depicting mortality before 28 days postnatal age, before 36 weeks postmenstrual age and before discharge, the latter was removed, since the data were exactly the same. In conclusion, there was no significant difference observed in mortality between the two groups; death occurred in 19 of 136 infants (14.0%) and in 25 of 137 (18.2%), respectively (absolute risk difference, -4.3 percentage points; two-sided 95% CI, -13.0 to 4.4).
In response to your other comments we would like to respond as following.
Paracetamol was given to 34 of 136 patients (25.0%) in the expectant management group, as compared to 52 of 137 patients (38.0%) in the early ibuprofen treatment group. So paracetamol was prescribed less frequently in the expectant management group. As the dosage used was the ‘normal’ analgesic dosage (20-40 mg/kg/day), which is lower than the dosage of 60 mg/kg/day investigated in randomised trials on PDA closure (2, 3), this is unlikely to have influenced our findings.
You questioned about the use of diuretics in this study. As described in the study protocol (4), patients randomised to the expectative management arm will not receive COXi, including for indications other than closure of the DA. No (additional) putative interventions to prevent or treat a PDA, for example fluid restriction or diuretics for that purpose only, are allowed. There was no significant difference in the use of diuretics between the study groups. The observed use of diuretics in this study population of extreme preterm infants is similar to published data. (5-7)
In your opinion the study was “drastically underpowered”. As mentioned in the paper as one of the limitations, ‘enrollment was stopped after only 48% of the planned sample size had undergone randomization’. However, this does not mean that the study was underpowered, let alone ‘drastically’ underpowered. The results are very clear and speak for themselves. With an absolute risk difference of -17.2 percentage points with a one-sided 95%-CI of -7.4 percentage points it can be concluded that expectant management for PDA in extremely premature infants was noninferior to early ibuprofen treatment with respect to necrotizing enterocolitis, bronchopulmonary dysplasia, or death at 36 weeks’ postmenstrual age.
Referring to the Figure in the blog of possible outcomes of a non-inferiority trial, our results are even consistent with the upper outcome, noninferiority and superiority, since the upper boundary of the one-sided 95% confidence interval didn’t overlap the value of 0%.
We’re sorry to hear that you consider our composite primary outcome as “weird”. We fully understand and acknowledge that parents of the infants that are treated on our NICU’s are not only interested in that outcome at that specific age. That’s why we’re very pleased to have a very inspiring and good collaboration with Care4Neo, the Dutch organisation representing the interests of preterm and newborn infants and their families. Care4Neo was also involved in this study and the publication.
As depicted in Table 3 and S7 there were no differences between the groups for need for supplemental oxygen, and length of hospitalisation. There was a slight, significant difference in the time to full enteral feeding, which was shorter in the expectant management group (Table 3).
Every study will raise additional research questions, and the follow-up of the BeNeDuctus study population is of major importance. As published in the study protocol, all patients are evaluated at a corrected age of 24 months.
Regarding your opinion about ‘many other rather strange choices in data presentation’, such as West syndrome and wrist abscess in the list of adverse events, we would like to say that all reported adverse events are summarised in Table 4. It would be very negligent to exclude West syndrome, when this has been reported by one of the centres as an adverse event.
Hopefully we have clarified important issues to your satisfaction. In our opinion, the results are really important and the suggestion of potential harm of Ibuprofen should be taken seriously and investigated in more detail. Unfortunately, not all relevant results are immediately known and we would like to invite you to take notice of subsequent publications regarding the BeNeDuctus Trial in the near future.
Tim Hundscheid and Willem P. de Boode
References
Hundscheid T, Onland W, Kooi EMW, Vijlbrief DC, de Vries WB, Dijkman KP, et al. Expectant Management or Early Ibuprofen for Patent Ductus Arteriosus. N Engl J Med. 2022.
Harkin P, Harma A, Aikio O, Valkama M, Leskinen M, Saarela T, et al. Paracetamol Accelerates Closure of the Ductus Arteriosus after Premature Birth: A Randomized Trial. J Pediatr. 2016;177:72-7 e2.
Ohlsson A, Shah PS. Paracetamol (acetaminophen) for patent ductus arteriosus in preterm or low birth weight infants. Cochrane Database Syst Rev. 2020;1(1):CD010061.
Hundscheid T, Onland W, van Overmeire B, Dijk P, van Kaam A, Dijkman KP, et al. Early treatment versus expectative management of patent ductus arteriosus in preterm infants: a multicentre, randomised, non-inferiority trial in Europe (BeNeDuctus trial). BMC Pediatr. 2018;18(1):262.
Hagadorn JI, Sanders MR, Staves C, Herson VC, Daigle K. Diuretics for very low birth weight infants in the first 28 days: a survey of the U.S. neonatologists. J Perinatol. 2011;31(10):677-81.
Gouyon B, Martin-Mons S, Iacobelli S, Razafimahefa H, Kermorvant-Duchemin E, Brat R, et al. Characteristics of prescription in 29 Level 3 Neonatal Wards over a 2-year period (2017-2018). An inventory for future research. PLoS One. 2019;14(9):e0222667.
Guignard JP, Iacobelli S. Use of diuretics in the neonatal period. Pediatr Nephrol. 2021;36(9):2687-95.
The lack of mortality between 36 weeks and discharge is reassuring, but surely it should have been in the initial publication? This is not, unfortunately infrequent, there are a few other trials where it has been difficult to find the mortality after 36 weeks, even though such deaths are uncommon, they can make a difference to the interpretation of the results. The stop-BPD trial for example, had “statistically significant” difference in mortality at 36 weeks, but not at discharge, and not at 2 years follow-up, there were actually in that study 17 deaths between 36 weeks and discharge (8 vs 9 in the 2 groups) so the p-value was just over .05 at discharge, after being just under .05 at 36 weeks. Another illustration of why we should stop using simple p-value thresholds to decide if something is real or not!
It also is not necessary to adjudicate all outcomes at the same moment! Even if BPD is decided at 36 weeks, mortality before discharge can still be the primary survival outcome, and even if some babies with BPD die between 36 weeks and discharge it is simple to count them as a death, and as a BPD outcome.
Which brings me to the issue about my calling ‘death or BPD at 36 weeks’ a “weird” primary outcome. What I meant is that I don’t decide whether to give a medication or not based on what the baby will be doing at 36 weeks. The decision is based on whether the baby is more or less likely to survive, and, if they survive, how they will evolve over their time in hospital, and after.
Determining the severity of lung injury by need for oxygen (or respiratory support) at 36 weeks is a very common practice in randomized trials and in epidemiologic studies of preterm babies. This comment is, therefore, not really directed at the investigators of the Beneductus trial alone, but at all of us as we go forward in neonatal research. I note again, that even though the proportion of infants with “BPD” was lower in the expectant treatment group, they only had (as a median) 1 day less oxygen treatment than the ibuprofen group. The big difference in BPD (33% vs 51%) despite almost no change in median duration of oxygen therapy suggests strongly that a lot of the babies labelled as “BPD” came out of oxygen very soon after 36 weeks.
In day to day practice, I don’t actually care if a baby comes out of oxygen before or after 36 weeks, and parents don’t care either (we asked them: article in submission, I will blog about that study when it is finally accepted). What matters to parents is whether the baby goes home in oxygen, whether their discharge is delayed by respiratory concerns, whether they sleep and eat normally after discharge, whether they have to make multiple hospital or doctor’s office visits for their respiratory problems. I was very happy to see that the Care4Neo group was involved in the study, it is essential for the future that such groups are involved, and in particular that they are involved in the development of our primary outcome variables.
I recognize of course, that we need interim outcomes, and that all trials cannot be done with the primary outcome being respiratory symptoms up to adolescence!! Surely we should be analyzing the impacts of our interventions on outcomes which are important to babies and their families, those interim outcomes should be determined with parent groups. They might well be interested in home O2 therapy, delayed discharge for respiratory reasons, and home gavage, perhaps. Oxygen need at 36 weeks has very limited predictive capacity for future respiratory health. (Barrington KJ, et al. Respiratory outcomes in preterm babies, is bronchopulmonary dysplasia important? Acta Paediatr. 2022).
I think the Beneductus trial does show that there were no clear benefits of treating the PDA, and that a dedicated group of researchers can construct a trial in which there is a group in which almost no-one receives a cox inhibitor. The trial does suggest that such an approach is not worse than early routine ibuprofen for a large PDA with left to right shunt, but there are still concerns about power, despite what is written in the comment: post-hoc power analysis is useful, but after early termination of a trial must be interpreted carefully. At least in this instance the trial was not terminated after examining the data, which is always dangerous, but because of enrolment difficulties and funding issues.
The follow up of the Beneductus trial will be very informative, I would be surprised if there were any neurological or developmental differences between groups, so other health outcomes will be very interesting. If they show no advantage on longer term respiratory health (addressing outcomes of importance to families) then it would suggest that it doesn’t matter whether you treat the PDA with cox inhibitors or not.
Which makes me wonder if there might be still be a role for the drugs in babies at very high risk of pulmonary haemorrhage. A few years ago we introduced a protocol of very early screening and treatment of PDA, largely based on the trial of Martin Kluckow (Kluckow M, et al. A randomised placebo-controlled trial of early treatment of the patent ductus arteriosus. Arch Dis Child Fetal Neonatal Ed. 2014;99(2):F99-F104) showing less hemorrhage with treating a group who were at high risk of hemorrhage, based on having a PDA diameter above the 50th percentile for their postnatal age. Which means, of course that about half of the screened babies are eligible. In the Beneductus trial, using a PDA diameter of 1.5mm, about 2/3 of screened babies were eligible. I wonder if we can refine the criteria and target a subgroup where treatment will lead to the advantage of fewer pulmonary haemorrhages, which, even if we cannot prove an improvement in long term outcomes, is still something that I would like to avoid!
Babies of under 28 weeks gestation were randomized at 24 to 72 hours of age if an echocardiogram showed a PDA of over 1.5 mm diameter, and a majority left to right shunt. They either had expectant management (no pharmacological or surgical PDA closure, which was followed by all except 1 of the 136 in this group) or a course of ibuprofen, followed by a second, and then possibly a third and fourth courses, with an option for ligation under limited circumstances, n=137 in the intervention group.
My first question is whether infants in the PDA closure group were more likely to survive or not, and I don’t know! There is nowhere I can find in the document or supplemental material the mortality before discharge in the 2 groups. The intervention group had much higher mortality before 28 days, 18% vs 9.6%, and somewhat higher at 36 weeks, 18% vs 14%, but death before discharge (which is listed as an outcome in the supplement) is nowhere reported.
I don’t know about you, but I can’t interpret, or care very much, about any of the other outcomes if I don’t even know if survival is different.
It looks like there was about the same NEC in the 2 groups, and more moderate to severe BPD in the intervention group, so there was more of the composite outcome “death prior to 36 weeks, or mod/severe BPD or NEC (stage 2 or 3)” in the ibuprofen group, 64% vs 46%.
There are a couple of other issues that are not clearly addressed in the publication, a substantial proportion of the babies received paracetamol (acetaminophen to the North Americans), 38 % of the ibuprofen group and 25% of the expectant group, but whether this was for analgesia or as therapy to close the PDA is not mentioned. There were also a lot of babies who received diuretics, over 40% in each group, but for what indication, at what time, and for how long isn’t clear.
The study was stopped before halfway, so it was drastically underpowered to answer the weird question of what ductal treatment with cox inhibitors does to outcomes up to 36 weeks. I say a weird question because it is one that I don’t really care about at all! Were the babies more or less likely to survive to go home if they were treated? Did they have worse pulmonary outcomes, i.e. home oxygen, or prolonged hospitalisation for pulmonary reasons, or poor feeding because of persistent respiratory distress?
For the outcome that was reported, expectant management was “not inferior” to intervention. In other words if you stop looking at the babies after 36 weeks PMA, then leaving the PDA alone is not worse than treating it, but you have to guess about what happens after 36 weeks.
IF there were few deaths between 36 weeks and discharge, then expectant management appears preferable, but the relevance of the parts of the primary outcome are questionable: at 36 weeks there were more intervention group babies who needed oxygen, but the median duration of oxygen therapy (40 vs 41 days) and of respiratory support (55 vs 56 days) were almost identical between the 2 groups. Why is there no report about outcomes that matter? How many went home on oxygen? Was hospitalisation prolonged in the group with more BPD? Were the same number of babies feeding orally when they were discharged? These are things which have impacts on families, far more than the diagnostic label of BPD, and none of them are mentioned in the publication. They have made many other rather strange choices in data presentation, for example, I know that one of the babies developed West syndrome (which was likely diagnosed after 36 weeks!), and, from the table of adverse effects, I know that one baby in each group had a wrist abscess, despite not knowing the survival!!
There is 2 year follow up planned, so hopefully we will eventually discover the survival rates in the 2 groups in the trial.
Another trial published this year had some similarities (Potsiurko S, et al. Randomized Noninferiority Trial of Expectant Management versus Early Treatment of Patent Ductus Arteriosus in Preterm Infants. Am J Perinatol. 2022). In a single centre trial from Lviv, Ukraine, VLBW infants of under 32 weeks gestation were randomized at less than 72 hours of age if their PDA diameter was over 1.5 mm. They received either attempted closure (with either rectal ibuprofen or intravenous acetaminophen, n=104, the choice of rectal ibu or IV acéta was apparently randomized), or expectant treatment, n=104. with 8 of the expectant group having “rescue” PDA treatment at about 7 days of age.
These were, on average more mature babies, with a lower incidence of BPD, and there were no important differences in either survival to discharge (about 80% in each group), or the need for oxygen at 36 weeks (28% with active treatment vs 22% with an expectant approach). The power of the trial for excluding even major differences in outcomes is, of course, limited, with this sample size.
Most importantly however, I think there has been a huge failure of peer review for this trial, which is written in some parts as, and stated in the title to be, a non-inferiority trial, but it is no such thing. It is a comparative trial that did not show a difference between the groups, that is NOT the same as showing non-inferiority!
The sample size calculation is very strange to start off with, it starts by assuming that there is an outcome difference in favour of expectant treatment (death or BPD of 35% compared to 55%), and then calculating for an alpha of 0.01 and a power of 90% that they would not show a difference of greater than 10% beyond this difference. Which gave them a very small sample size for a non-inferiority trial, of 84 per group.
That really isn’t what is meant by a non-inferiority trial! The point of a non-inferiority trial is to show that the two approaches are not very much different, and that the experimental arm is not worse than the usual care arm, by an acceptably small amount, with a certain confidence. So to calculate the sample size you should start with the assumption that the two treatments have about the same outcome, and then calculate how much worse the new treatment could be without exceeding your “acceptably small amount”. If you use the non-inferiority margin that these authors have claimed in the methods of 10%, and that the incidence of “death or BPD” with standard care is about 40%, then with an alpha of 0.05 and a power of 80% (which are much more conservative than the calculations these authors have used) the sample size required is 297 PER GROUP.
There is no mention in the results of the confidence interval of the difference between groups, which is essential to know if the non-inferiority margins were crossed or not. What I think they have actually shown is NOT non-inferiority, I have tried to visualize the results with the figure below, which shows the event difference between groups (death or BPD was 0.45 vs 0.39, or an event difference of 0.058, with 95% CI of -0.085 to 0.195, which overlaps both no difference and the non-inferiority margin of a 10% difference, and is therefore an inconclusive result.
You cannot say, based on this result, that expectant treatment is non-inferior to active PDA closure.
As a reminder, here are the possible outcomes of a non-inferiority trial, in a similar graphic form (from Mauri L, D’Agostino RB, Sr. Challenges in the Design and Interpretation of Noninferiority Trials. N Engl J Med. 2017;377(14):1357-67)
A large proportion of the placebo babies received open-label ibuprofen after the initial course (62%, compared to 17% of the ibuprofen group), so the study is not very informative for other clinical outcomes, which really weren’t very much different between groups even though the PDA was more likely to be closed on day 3 and on day 14 in the ibuprofen group.
The primary outcome of this trial was the diagnosis of cerebral palsy at 2 years of corrected age. Survival to 2 years was a little more frequent in the placebo group, and CP was a little less frequent, but both differences could easily have been due to chance.
There are other trials of treating the PDA compared to a conservative approach, hopefully with designs which will inform us for the future which approach is preferable, based on clinically important effects such as survival to discharge and other outcomes which are of importance to babies and their families.
Still on my trip to Ecuador, where I saw my first ever Canada Warbler. I know it is just a name, but it was most pleasing to find this bird, which does indeed migrate to Canada each year but I had never encountered, and passes the Canadian winter here.
I have completely changed the design of my other blog keithbarrington.com hopefully it is easier to navigate and see anything that might interest you.
A break in blogging for a couple of weeks, while I take a trip to Ecuador, with the main purpose of birdwatching and photography. Ecuador is the country with the greatest avian biodiversity in the world, and with and amazingly knowledgeable guide, I’m having a great experience.
I am blogging about the trip as I go, mostly in order to post photos. Please visit https://keithbarrington.com/, in addition to the home page, I will add various other pages of photos, for now, there is just the page “All the Tanagers of the Trip” which you can find by clicking a link at the top of the page.
Neonatal Research 